For the smallest recipients of life-saving bone marrow transplants, the gravest threats do not always announce themselves with fever or crisis; some arrive as a whisper. A new study from China reports that a subtle but persistent decline in a child’s breathing capacity, a drop of just ten percent in the volume of air exhaled, can foreshadow bronchiolitis obliterans syndrome, one of the most feared and lethal complications of allogeneic hematopoietic stem cell transplantation. By systematically tracking lung function in 201 children over nearly a decade, researchers identified a cluster of early warning signals that flagged the syndrome months before it would otherwise have been diagnosed. The findings, published on 29 August 2026 in Annals of Hematology, suggest that routine spirometry, already performed in many transplant centers, may contain information of far greater value than previously appreciated, provided clinicians learn to read it not as an isolated number but as a trajectory, and to act on the change before scar tissue locks the airways shut.

Allogeneic hematopoietic stem cell transplantation, abbreviated allo-HSCT in the shorthand of hematology, replaces a child’s diseased blood-forming system, destroyed by intensive chemotherapy for leukemia or failing because of inherited immune deficiencies or bone marrow failure disorders, with stem cells donated by a matched donor, which rebuild the marrow from scratch. The same donor immune system that rescues the child, however, can turn against the recipient’s own tissues, a complication known as graft-versus-host disease. Donor T cells commonly attack the skin, gut, and liver, but when the immune assault targets the lung’s smallest airways, it can set off obliterative bronchiolitis: an inflammatory process in which the bronchioles, airways narrower than a drinking straw, become swollen, scarred, and progressively constricted by fibrous tissue. The resulting clinical syndrome, bronchiolitis obliterans syndrome (BOS), is defined not by lung biopsy but by an otherwise unexplained, progressive fall in airflow measured on breathing tests. Because the fibrosis is essentially irreversible, scarred bronchioles do not reopen, and the syndrome carries a high mortality that is driven, in large part, by how late it is caught.

The central clinical frustration has long been timing. BOS is typically recognized only after lung function has deteriorated past a fixed threshold and alternative explanations, such as infection, have been excluded, a framework that almost guarantees a late diagnosis. Conventional surveillance has tended to be static: a snapshot of lung function at a given visit, compared against broad population norms rather than the child’s own baseline, and interpreted in isolation from the preceding weeks and months. Yet the disease itself is dynamic, unfolding gradually as immune-mediated injury accumulates in the small airways and airflow quietly erodes. For children the stakes of that framing are even higher, because growing lungs should push spirometry values upward as a child gets taller; a decline that would be unremarkable between two adults may therefore be a genuine red flag in a child. The team set out to test a deceptively simple idea: whether shifting the analytical frame from static thresholds to longitudinal change could convert ordinary follow-up testing into a true early-warning system.

The study was a single-center, retrospective cohort analysis carried out by a team at the Department of Pediatric Hematology of the First Affiliated Hospital of Shandong First Medical University and Shandong Provincial Qianfoshan Hospital in Jinan, with hematologist Hongmei Wang as corresponding author. The investigators reviewed the records of 201 children who underwent allo-HSCT between January 2015 and December 2024, dividing them into 29 patients who developed BOS and 172 who did not. Across the entire cohort, the cumulative incidence of the syndrome was 6.9 percent, a figure that appears modest but represents a substantial burden given the severity of the outcome. To identify which characteristics independently predicted BOS rather than merely accompanying it, the researchers applied multivariate Cox proportional hazards regression, a statistical technique that estimates the effect of each candidate risk factor on the timing of disease onset while mathematically holding every other variable in the model constant, thereby disentangling genuine predictors from statistical bystanders.

Four factors emerged as independent predictors, each expressed as a hazard ratio that quantifies how strongly the condition multiplies the instantaneous risk of developing BOS. The most powerful was severe acute graft-versus-host disease: children who experienced grade III–IV aGVHD faced a hazard ratio of 15.12, meaning that at any given moment their risk of progressing to the syndrome was roughly fifteen times that of children who had escaped severe aGVHD. Chronic graft-versus-host disease multiplied the hazard 6.94-fold, and the appearance of a dry, non-productive cough, a symptom easily dismissed in a child recovering from transplant but one that likely reflects immune injury irritating the small airways, raised it 11.77-fold. Finally, an early post-transplant decline of ten percent or more in the FEV1/FVC ratio, a key spirometric index of airway obstruction, carried a hazard ratio of 8.70. Together, the four predictors sketch a coherent biological narrative in which uncontrolled donor-versus-host immunity first announces itself as symptoms and measurable airflow change, then hardens into obliterative bronchiolitis.

The technical heart of the study lies in two numbers that children undergoing transplantation already generate routinely. Forced vital capacity, or FVC, is the total volume of air a child can forcibly exhale after a maximal inhalation; forced expiratory volume in one second, or FEV1, is how much of that air leaves the lungs within the first second. Their ratio, FEV1/FVC, characteristically falls in obstructive airway diseases, because narrowed airways slow the speed at which air can be expelled. The team’s at-risk signal was defined as a relative FEV1 decline of at least ten percent from the child’s own post-transplant baseline, a criterion corresponding to the internationally recognized pre-clinical staging category known as BOS stage 0p. Because the benchmark is each patient’s personal baseline rather than a population average, the signal automatically adjusts for age, height, and growth, making it unusually well suited to a pediatric population in which absolute lung-function values are constantly moving targets and a single normal-looking result can mask an ominous downward trend.

When the researchers assessed how this signal performed, the result was striking. Children whose relative FEV1 decline crossed the ten-percent threshold reached a BOS diagnosis a median of 176 days earlier than they otherwise would have, nearly six months of additional time in which therapies could, in principle, be escalated before irreversible scarring took hold. The signal’s negative predictive value, the probability that a child without the flag is genuinely free of the disease, was 85.0 percent, offering considerable reassurance when the signal is absent. Its positive predictive value, however, was low: many children flagged by the ten-percent decline never went on to develop BOS. The authors are explicit about this trade-off, noting that a fall in FEV1 alone cannot justify a diagnosis or aggressive treatment changes; rather, the low PPV means the signal must be integrated with comprehensive evaluation, including symptoms, imaging, infection work-ups, and graft-versus-host assessment, before any clinical decision follows. The flag, in other words, opens an investigation rather than closing a case.

Perhaps the most conceptually provocative finding concerns the mismatch between what imaging and lung function each reveal. The team compared longitudinal trajectories of pulmonary function against scoring of high-resolution computed tomography, the fine-slice CT technique capable of resolving the bronchioles and lung architecture in fine detail, in which obliterative small-airway disease appears as air trapping, patchy mosaic attenuation of lung density, and progressive airway dilation. They documented a significant radio-functional discordance: in affected children, HRCT scores worsened even while FEV1 remained apparently stable, a divergence that reached statistical significance at P < 0.001. The implication is unsettling, because it means structural damage in the small airways can accumulate silently, outpacing the functional tests meant to detect it; conversely, a reassuring spirometry result cannot be taken as proof of a healthy lung in a transplanted child. The authors argue that this discordance underscores the need for combined functional and imaging surveillance, with each modality compensating for the blind spots of the other.

Translated into practice, the study proposes a two-tier strategy for post-transplant pulmonary care in children. At the population level, the identified risk factors allow prognostic stratification: a child with a history of grade III–IV acute graft-versus-host disease, chronic graft-versus-host disease, or an emerging dry cough belongs to a high-risk group in whom pulmonary monitoring could reasonably be intensified and the threshold for HRCT imaging lowered. At the individual level, serial spirometry interpreted as a trajectory rather than a series of disconnected points functions as a sensitive tripwire. The authors emphasize that dynamic FEV1 monitoring served in their cohort as both an early warning signal and a prognostic marker, while carefully framing it as one component of a comprehensive evaluation rather than a standalone diagnostic test. Even with that caveat, the ability to shift a diagnosis forward by nearly six months carries real weight in a disease whose damage is cumulative and permanent, and in which every week of untreated inflammation risks converting reversible injury into fixed, untreatable obstruction.

The researchers themselves temper the enthusiasm. The study is a single-center, retrospective analysis with a limited number of BOS cases, 29 children in total, and the authors caution that, given its exploratory nature, all findings should be interpreted with care. Retrospective designs inherit the blind spots of the records that feed them, very young children often cannot perform reliable spirometry, and the hazard ratios await confirmation in larger, prospective, multicenter cohorts before the ten-percent rule is written into guidelines. Still, the direction of travel is clear. As survival after pediatric stem cell transplantation continues to improve, the field’s attention is shifting from keeping children alive to protecting the quality of the lives saved, and the lungs have become a central battleground in that shift. For the roughly one child in fifteen who would otherwise face bronchiolitis obliterans syndrome, an ordinary breathing test read correctly, and read early, may prove to be the difference between a progressive lung disease caught too late and one intercepted in time.